The gas phase reaction of ammonia and gaseous elemental fluorine can produce nitrogen trifluoride. Reaction 1 illustrates the desired gas phase NF3 production reaction.3F2(g)+NH3(g)→NF3(g)+3HF(g)(ΔH=−904KJ/g mole NF3)  Reaction 1                 wherein (g) denotes the gas phase. A solid catalyst is often used to lower the required operating temperature, which increases the NF3 yield. However, it is very difficult to control the reactor temperature due to the highly exothermic nature of Reaction 1. As a result, the gas phase ammonia and fluorine reaction produces substantial quantities of HF, N2, N2F2, and NH4F, with NF3 yields typically substantially less than ten percent.        
U.S. Pat. No. 4,091,081 teaches a process that produces much higher nitrogen trifluoride [NF3] yields (approaching 60%) by contacting a molten ammonium acid fluoride [NH4F(HF)x] with gaseous fluorine [F2] and ammonia [NH3]. U.S. Pat. No. 5,637,285 to Coronell et al. describes a similar process, wherein the F2-to-NF3 conversion is further increased to greater than ninety percent by imputing a large amount of mechanical energy for mixing the reactants and by using an ammonium acid fluoride melt having a HF/NH3 molar ratio greater than 2.55. The Coronell patent teaches that improved NF3 yields are achieved with mechanical energy inputs greater than 1,000 watts per cubic meter, preferably at or above 5,000 watts per cubic meter, most preferably at or above 35,000 watts per cubic meter. The Coronell patent utilizes a stirrer or turbine, such as a flat blade turbine, to input the mechanical energy.
However, inputting such large amounts of power using a stirrer or turbine poses reactor reliability problems. Typically, the mixing turbines that are used in this type of application are constructed of a metal, such as monel or nickel, coated with a metal fluoride passivating layer. The passivating layer is typically applied by contacting the metal turbine with a fluorine rich atmosphere. The passivating layer significantly reduces oxidation of the turbine substrate. However, the high power inputs suggested by the Coronell patent produce high sheer rates that can remove the passivating layer and expose the underlying turbine substrate to fluorine, thereby accelerating the rate of corrosion, particularly at the tip of the mixing impeller. In turn, the corrosion leads to excessive mixer shaft vibration and premature mixer shaft seal failure. Even if very diligent maintenance procedures essentially eliminate shaft vibration, the combination of a high-speed rotating seal and a corrosive fluorine and hydrogen fluoride atmosphere can lead to reliability problems. In addition, high mechanical energy input into a reactor, via a flat bladed turbine, can lead to a highly back-mixed reaction volume with essentially uniform operating conditions. In this case, there is no opportunity to optimize the local reactor operating conditions. Therefore, there remains a need in the art for a method and apparatus to efficiently and reliably contact gaseous fluorine with NH4F(HF)x solution to produce nitrogen trifluoride.